Share via

How to use Integrated Vectorization option with Onelake(fabric) files

Ashish Kalsi 20 Reputation points
2024-07-22T05:33:18.0866667+00:00

So, I am trying to index my data from one lake files, using Lakehouse (fabrics). the documentation says either the URL or workspace id and Lakehouse id we can use to proceed further to create index.
I have used workspace id from the URL from the properties option.
User's image

so after.com/ the string i have used to apply as workspace id and lakehouse id I have already. So i used the wizard Import and Vectorize and ended up with creating indexer,index,skill set - that auto created by wizard . Now the error part is at my lakehouse there is data stored in form of mutliple tables , multples rows and columns as
q2

I wants to ingest data completely into index and It should work , I tried ended up getting this only ,q4

why I am not able to query ,
s3

Azure AI Search
Azure AI Search
An Azure search service with built-in artificial intelligence capabilities that enrich information to help identify and explore relevant content at scale.
0 comments

1 answer

Sort by: Most helpful
Most helpful Newest Oldest
  1. Amira Bedhiafi 39,106 Reputation points Volunteer Moderator
    2025-10-16T19:30:47.3733333+00:00

    Hello Ashish !

    Thank you for posting on Microsoft Learn Q&A.

    Integrated Vectorization in Azure AI Search is designed for documents in OneLake Files (PDF, DOCX, TXT, HTML, images to OCR...).

    Your lakehouse show data under tables and the wizard won’t pull rows from lakehouse tables and turn them into content automatically, so your indexer finishes with 0 documents.

    You can put your documents under lakehouse then files or add a OneLake shortcut there.

    In the wizard, choose OneLake (Fabric) and browse to the files path then keep integrated vectorization on so it will extract text then embed then store vectors.

    Then give your search service managed identity reader on the Fabric workspace or the Lakehouse item and if you use a user assigned MI, grant that principal access.

    You can also use text-embedding-3-small/large instead of text-embedding-ada-002.

    0 comments

Your answer

Answers can be marked as 'Accepted' by the question author and 'Recommended' by moderators, which helps users know the answer solved the author's problem.